The “symphonic” nature of natural systems

A system can be defined as an interacting set of parts. There are, of course, many kinds of system. I sometimes like to draw a parallel between natural systems and symphonic music (another kind of system), as a means of conveying certain ideas about what we gain from a multivariate approach to ecology. In symphonic music, a composition relies on the participation and interplay of many instruments. As we listen to an orchestra (preferably live for greatest effect) we gain an entirely different experience compared with the effects of listening to the instruments individually. The same is true with the study of natural systems. The preceding chapters have sought to convey through examples what can be gained from achieving a multivariate perspective. While descriptive multivariate methods have been increasingly employed by natural scientists, the study of hypotheses has largely relied on univariate methods. The result has been an emphasis on individual processes, with little opportunity to consider their role in the system. My own earlier experience with the study of competitive interactions convinced me of the value of building and evaluating multivariate hypotheses, where the importance and impact of individual processes can be seen in some context. It is, of course, true that the simultaneous study of all parts of a system is not feasible. However, it would seem that our understanding of the observed variation in a system property of key interest can often be improved without resorting to extremely complex models.